Appendix: Supporting Information for “tree-based Models for Political Science Data”

POLITICAL SCIENCE DATA” This supporting information (SI) appendix first provides additional details on the recursive binary splitting algorithm used to build individual trees. Second, it presents an additional discussion as to how these methods handle categorical outcomes. Third, it briefly provides additional discussion of k-fold cross validation. Fourth, it provides a more detailed discussion of several practical issues surrounding estimating and interpreting tree models. Finally, it presents additional details about the three applied examples in the main text.